We collected data at the volume-level, not at the title-level. Addressing individual volumes allowed the student workers’ efforts to be as simple and quick as possible. In retrospect, we recognize this was not a good choice; title-level data are more meaningful. For example, the thirty-four-volume Dictionary of Art had thirty-eight uses over five years, but thirteen of the individual volumes had no dots affixed. In this example, then, 34 percent of the books received a zero in the spreadsheet, even though the title was clearly used. It is apparent that reporting multivolume title use is problematic, with the data often indicating a higher level of non-use than is warranted. Should this study be replicated, this is one aspect that needs to change.
We realized, too late, that the one student reshelver we had for two consecutive years was not following the established procedures for affixing the dots in the books. Because in her first year she did an excellent job, we have postulated that she was assuming that heavily used books already had ten dots and thus failed to dot them. Whatever the reason, there was a notable drop in use of key reference items, books that are required and consistently heavily used in the Bible College and Seminary courses. Indeed, the entire count for that year was 30 percent lower than that of the previous year, and 43 percent lower than that of the following year. Other reference statistics as well as observations of reference staff convince us that the data provided through the dots for that year are unreliable; they are accurate in what they do report, but inaccurate in what they do not.
For Future Research
Is aversion of librarians to weeding as pervasive as the literature seems to indicate? It is mentioned often, but more as an aside than as an empirically verified fact.
What is happening with reference book use statistics as electronic resources abound? If, as one might suspect, use is dropping, what are the implications? Generally, online databases do not contain the content found in the high-quality volumes that populate academic reference collections. If print reference use is decreasing, can it be assumed that our students, faculty, and even librarians are less scholarly in their quests for information? What are the weeding implications if use is dropping precipitously?
Would a follow-up study be correctly considered a follow-up study when electronic access to materials might significantly impact the use statistics of the print reference collection?
Is it verifiably, consistently true that more books in the lean reference collection will be used than in the fat (unweeded) collection?
Conclusion
The GAF reference staff thought that with collecting data at the end of the five-year study they would have finished the work. They now realize that the study has provided the information needed to begin several tasks. First, a mission statement and collection management policies will be developed. Library literature provides many articles that can be useful in guiding these tasks.27 It is, of course, understood that, once written, these documents are designed for guidance and not only can, but should, be reviewed regularly and revised as necessary. They are intended to aid librarians as we seek to meet patron needs, and should be utilized only insofar as they fulfill that intent.
After the mission and policies have been developed, a plan for weeding must be developed. Vincent’s entertaining article provides excellent commentary regarding the extent to which reference weeding involves staff from throughout the library.28 Therefore, a plan that is developed in coordination with representatives from circulation and technical services is vital. Once the plan is in place, then both selection and deselection can happen within the parameters of the mission and policies.